Atomic Force Microscopy of Cell Growth and Division in Staphylococcus aureus

The growth and division of Staphylococcus aureus was monitoredby atomic force microscopy (AFM) and thin-section transmissionelectron microscopy (TEM). A good correlation of the structuralevents of division was found using the two microscopies, andAFM was able to provide new additional information. AFM wasperformed under water, ensuring that all structures were inthe hydrated condition. Sequential images on the same structurerevealed progressive changes to surfaces, suggesting the cellswere growing while images were being taken. Using AFM smalldepressions were seen around the septal annulus at the onsetof division that could be attributed to so-called murosomes(Giesbrecht et al., Arch. Microbiol. 141:315-324, 1985). Thenew cell wall formed from the cross wall (i.e., completed septum)after cell separation and possessed concentric surface ringsand a central depression; these structures could be correlatedto a midline of reactive material in the developing septum thatwas seen by TEM. The older wall, that which was not derivedfrom a newly formed cross wall, was partitioned into two differentsurface zones, smooth and gel-like zones, with different adhesiveproperties that could be attributed to cell wall turnover. Thenew and old wall topographies are equated to possible peptidoglycanarrangements, but no conclusion can be made regarding the planaror scaffolding models.

INTRODUCTION

Top

Introduction

The cell wall is an essential structure for almost all bacteria,functioning as a rigid exoskeleton for protection against bothmechanical and osmotic lysis and for contouring cellular shape(7). The cell wall is also important in controlling cellularinterfacial interactions with the environment, cell-substratumadhesion, and cell-cell aggregation (4, 12).

Over the past decade, much progress has been made in understandingthe surface ultrastructure of bacteria. Electron microscopy(EM) has long been recognized as a key technique in microbiologyto elucidate such surfaces (7-10), and the development of cryotechniqueshas just begun to allow high-resolution imaging of vitrifiedcell structures in conditions close to the native hydrated state(39, 31). Yet, such cryo-methods are technically demanding andpossible in only a few laboratories worldwide. Direct observationof a bacterium in aqueous solution is, however, difficult, makingdirect measurement of molecular interaction forces at the cellsurface rare (2, 3, 14, 15, 18-21, 30, 35, 38, 40, 41, 42).Since its discovery in 1986, atomic force microscopy (AFM) hasbecome an important tool for investigating biological materialssuch as tissues, yeasts, bacteria, and their components (1,18-20). By imaging bacteria in aqueous solution, AFM can providereal-time in situ quantitative morphological information aswell as measurement of the interaction forces between AFM tip(or modified probe) and cell surface.

However, there are a number of practical challenges that mustbe overcome when such measurements are attempted on bacteria.One problem is that the cells can be deformed under the loadingforce applied by the AFM tip. This leads to a reduction in lateralresolution and complicates the deconvolution of surface forcesand deformation in force curve analysis. Another problem isthe difficulty associated with sample preparation; it is crucialthat cells be well attached to a solid substrate. Without thisstrong attachment, cells can move in response to the lateralmovement of the tip, and can be pushed across the surface duringimaging. Since bacteria adhere to inanimate surfaces via weak-bondingforces that are readily broken by the tip movement, cells arefrequently knocked away before imaging. Therefore, various immobilizationprocedures have been developed, such as air drying, chemicalfixation, and immobilization in hydrated gels (18-20, 26, 42,44). Unfortunately, these methods can artificially alter bacterialsurfaces.

One way to solve this problem is to trap bacterial cells inporous polymer membranes (18-20, 38). Using this method, wehave employed contact-mode AFM to investigate real-time changesin the staphylococcal cell wall as cells grow and divide ingrowth medium. To assist interpretation of the resulting AFMimages, the data are compared to electron micrographs of thinsections from conventionally embedded staphylococci at similartime points of growth and division. Furthermore, AFM force curveswere recorded at different locations on the growing and dividingsurfaces of cells. Such curves provide valuable informationon the tip-surface interaction, which is sensitive to the chemicalnature of cell surface macromolecules.

MATERIALS AND METHODS

Top

Materials and Methods

Bacteria and growth conditions. Staphylococcus aureus D₂H is a representative staphylococcusand was used for study. It was maintained on Trypticase soyagar slants (Difco) until experimentation. A slant culture initiatedan overnight Trypticase soy broth culture that was used forthe mid-exponential growth phase (optical density at 600 nm= 0.2) used for study. All liquid cultures were grown at 37°C,whereas the AFM liquid cell was held at 20°C in an attemptto slow growth for better surface visualization.

Cell immobilization. A 10-ml cell suspension was gently sucked through an isoporepolycarbonate filter membrane with a pore size comparable tothe cell size (diameter, 1.2 µm). After filtration, thefilter was gently rinsed with deionized water to wash cellsof surface debris, turned upside down, and attached to a glasssubstrate using a small piece of double-sided adhesive tape.The mounted sample was then transferred into the AFM liquidcell containing Trypticase soy broth. All manipulations beforemicroscopy were rapidly accomplished so as to ensure cells remainedhydrated and were not subjected to high shear (which would dislodgethem). This protocol produced many filter pores whose ends wereplugged with a single cell. These staphylococci were so stronglyentwined with the polycarbonate meshwork of the filter thatcontact-mode AFM could be performed over a large part of theirprotruding surfaces.

AFM measurements. AFM images and force measurements were recorded in contact modeat room temperature using a Molecular Imaging Microscope. Imageswere recorded in both height and deflection modes. While heightimages provided quantitative information on sample surface topography,deflection images often exhibited higher contrast of morphologicaldetails. Imaging forces were kept between 0.5 and 1.0 nN, andscan rates were between 2 and 4 Hz. V-shaped cantilevers withoxide-sharpened Si₃N₄ tips (Veeco Co.) were used with springconstants of 8 ± 4 mN/m. Cantilever spring constantswere determined by measuring the tip deflections for known appliedforces as described by S. K. Jericho et al. (25). All measurementswere performed in growth medium, and all images are representativeof several different cells, each imaged at the same stage ofgrowth or division.

Transmission EM (TEM). As with AFM, bacteria were harvested at mid-exponential growth.They were then fixed via a glutaraldehyde-osmium tetroxide protocol,dehydrated using an ethanol-propylene oxide series, and embeddedin LR White (London Resins, Marivac, Nova Scotia) accordingto the method of Beveridge et al. (11). Thin sections were stainedwith uranyl acetate and lead citrate. Imaging was performedusing an FEI CM10 or an LEO 912AB device, both operating at100 kV under standard conditions.

RESULTS AND DISCUSSION

Top

Results and Discussion

General views of bacteria. TEM of thin sections revealed that the staphylococci were invarious states of growth and cell division. The cell wall wasan amorphous matrix ca. 25 nm thick and lay on top of the plasmamembrane (Fig. 1). No high-order structure was seen that couldbe attributed to molecular arrangements of peptidoglycan orsecondary polymeric or proteinaceous components, and no capsulewas distinguished. These images correlated well with low-resolutionAFM images of the bacteria; single spherical cells were seenstudding many of the filter membrane pores (Fig. 2), and thesespherical cells had rather featureless surfaces. To ensure littledeformation, AFM imaging forces of ca. 1 nN were used and imagesof the same area of each cell could be repeatedly obtained withoutdetaching the cell or altering surface morphology. The bacteriawere easily differentiated from the flat polymer surface ofthe filter (Fig. 2). The advantage of height imaging is thatit provides a quantitative measure of cell dimensions, layerthickness, and surface roughness while the bacterium is hydrated.Each entrapped cell showed an artifactual ridge at one edgeattributable to the initial contact between AFM tip and pore(Fig. 2).

View larger version (237K):
[in this window]
[in a new window]
FIG. 1. High-magnification TEM of a thin section of S. aureus D₂H showing the cell wall (CW) and plasma membrane (PM). Bar = 50 nm.

View larger version (91K):
[in this window]
[in a new window]
FIG. 2. Three-dimensional AFM image of two S. aureus bacteria trapped in pores of a polycarbonate filter membrane. The raised pore edge (black arrow) is an imaging artifact. The white arrow indicates the flat filter substrate. Image size, 4 by 4 µm; height of bacteria above the filter plane, 600 nm.

Features attributable to growth and division. One of the first observable features of growth and divisionamong staphylococci is the onset of septation, which, as seenin thin-section TEM, is a visible growth of wall material intothe cytoplasm (Fig. 3A). This septum will continue its inwardgrowth (like the closing of a camera aperture) until a cross-wallis formed that completely separates the two daughter cells (Fig.3B to D). At this point the cells remain joined together buteventually, through a process called splitting, separate (Fig.4A and B). During septation a small cleft is seen in the peripheryof the wall (Fig. 3C and D). This stage is the first topologicalfeature that is seen by AFM, where a shallow cleft is foundin the relatively smooth homogeneous surface at the midpointof a cell (Fig. 5A). A single x,z scan by the AFM tip at rightangles to the division plane (i.e., an AFM cross section) showsthis septal depression clearly (Fig. 5B is the scan outlinedby the solid line across the cell in Fig. 5A). As septationprogresses, the septum grows further inwards into the cytoplasmuntil the cross-wall is formed (Fig. 3D). The depth of the cleftwill depend on the extent to which septation has progressed.Since TEM and AFM images were taken on different cells and atdifferent stages in the division process, a comparison of thedepth of the clefts may not be meaningful.

View larger version (226K):
[in this window]
[in a new window]
FIG. 3. (A to D) Sequence of division seen in D₂H by thin-section TEM. (A) The septum has just been initiated and can be seen at the upper and lower periphery of the cell. The leading edge of each portion of the septum has a plasm membrane intrusion attached to it. (B and C) These are more advanced stages of septum in-growth, and membrane intrusions can still be seen at the leading edges. (D) Here, the septum is complete and bisects the cell to form (eventually) two daughter cells. (B and C) A dark midline of highly reactive dense staining material can be seen along the center of the septum. (C and D) The arrows point to the peripheral regions where the septum has commenced to split. Bars = 200 nm.

View larger version (112K):
[in this window]
[in a new window]
FIG. 4. (A) The daughter cells are actively splitting away from one another so that only approximately half of the septum remains intact. (B) Here, the cells have actually split away from one another, and now they each can be called a daughter cell. The arrows point to the new regions of the cell wall (approximately half of the total cell wall) that had before formed the septum. Bars = 200 nm.

View larger version (65K):
[in this window]
[in a new window]
FIG. 5. (A) Single bacterium immobilized in a filter pore. The septum furrow shows that the bacterium is in an early stage of cell division. (B) Cross-section along the line in image panel A. The arrow indicates the cleft position. Bar = 140 nm.

Throughout this process a darkly staining middle line (or midline;Giesbrecht et al. [23] refer to this as the "splitting system")is seen running down the center of the septum by thin-sectionTEM (Fig. 3B to D). Since staining for TEM requires the bindingof heavy metal stains to reactive sites within biomaterials,this midline is a region within the growing septum or cross-wallthat is distinctly different from the surrounding wall fabric(Fig. 6), containing more reactive sites (11). At the peripheryof the wall, at such advanced septation sites, the cell wallbecomes more indented as cell splitting is initiated; here,small fibers are revealed by thin-section TEM (Fig. 7; see theslightly darker strip that appears to span the cleft). A similardarkening is also visible in Fig. 3D (bottom cleft). This moreadvanced septation, where cell splitting begins, can also beseen as a series of regularly situated perforations by AFM observationsspaced approximately 100 nm apart (Fig. 8A). At higher magnification,these perforations were seen as holes approximately 50 to 60nm in diameter (Fig. 8D), as was confirmed by x,z scans (Fig.8E is the region in Fig. 8D outlined by the solid line). Correlationbetween the fibers seen by TEM (Fig. 7) and the holes seen byAFM (Fig. 8A to D) suggest they are one and the same. Afterapproximately 1 h, AFM on the same dividing cell revealed thatthe holes had enlarged until they merged together to form asingle larger perforation (Fig. 8B and C). Two explanationsare possible for this occurrence: (i) this is a real-time progressionof cell splitting where autolysins are systematically cuttingbonds in the peptidoglycan to ensure eventual daughter cellseparation, or (ii) this is a tip-induced deformation that enlargesthese delicate sites. If it is a deformation, this particularsite is clearly more deformable than surrounding wall regions(which also implies autolysin activity). Interestingly, theheavily stained midline seen by TEM just below this region (Fig.3B to D, 6, and 7) indicates a concentration of reactive sites,thereby implying that bonds have also been broken here. Thisis where the eventual splitting between cells will eventuallyprogress to. More advanced division stages clearly shows thisprogression in thin section where the daughter cells are almostseparated (Fig. 4A).

View larger version (227K):
[in this window]
[in a new window]
FIG. 6. A high magnification of a developing septum for comparison with Fig. 3B so that the darkly stained midline can be seen. Bar = 100 nm.

View larger version (225K):
[in this window]
[in a new window]
FIG. 7. A high magnification of a more-advanced septum showing the initial phase of splitting. The two arrows point towards small fibers remaining associated with each wall surface after splitting has occurred, thereby forming small holes in the septal fabric. These could be analogous to the small holes also seen by AFM (Fig. 8). Bar = 50 nm.

View larger version (83K):
[in this window]
[in a new window]
FIG. 8. (A to C) AFM deflection images showing advanced stages of septation in a cell. (A) The furrow is perforated by a series of equally spaced holes. (B and C) Hole enlargement over about a 1-h period at room temperature. (D) AFM height image of the cell in panel A. (E) Cross section through the perforation holes parallel to the line shown in panel D. The 50- to 60-nm-diameter holes have a center-to-center spacing of about 100 nm. (A) Bar = 220 nm; (B and C) bar = 120 nm.

Eventually, separation is complete, and what used to be partof the septum now becomes newly formed cell wall (Fig. 4B).As seen by the deflection mode of AFM, this new wall has a highdegree of infrastructure that is not apparent by thin-sectionTEM. Circular or concentric rings are seen enclosing a centraldepression (Fig. 9A), which are similar to the circular patternsseen by Amako et al. via freeze-etching TEM and scanning EM(SEM) (4, 5), Giesbrecht et al. via thin-sectioning TEM (22,23), and Beveridge et al. via negative-staining TEM (8). Anx,z scan through the central depression showed it to be at least30 nm deep (Fig. 9B shows a height mode AFM image, and Fig.9C is the x,z scan of the solid line seen in Fig. 9B). Examinationof radial sections, such as that in Fig. 9B, taken in severaldirections around the central depression, showed a spectrumof spacings between ring structures. A histogram of the spacingsmeasured on several cells is shown in Fig. 9D; spacings of 13to 25 nm were common, but larger spacings also occurred. Highermagnifications verified the concentricity of the rings (Fig.10A), and x,z scans revealed that the height of the rings increasedas they approached the central depression (from $~$ 1 to $~$ 2 nm; Fig.10B follows the solid line from left to right). Interestingly,each ring also possessed its own topography, with a height andamplitude of 1 to 2 nm (Fig. 10B).

View larger version (96K):
[in this window]
[in a new window]
FIG. 9. (A and B) AFM deflection and height images, respectively, of a newly formed cell wall (4, 5, 8, 22, 23), showing concentric rings that surround a central depression. (C) Cross section along the line indicated in panel B, showing a 30-nm-deep depression at the center. (D) Histogram of the ring periods obtained from many radial cross-sections on image (B). Spacings of 13 to 25 nm were common, but larger spacings were also observed. Bar = 50 nm.

View larger version (87K):
[in this window]
[in a new window]
FIG. 10. (A)AFM height image of a new cell wall showing a portion of the ring structure surrounding the central depression. Image size = 600 by 600 nm. (B) Cross section along the line shown in panel A. The corrugation amplitude increases towards the central depression.

The concentric ring structures were not visible on older regionsof the cell walls. High-resolution AFM imaging of these olderareas, however, showed what appeared to be a network of fiberswith a large number of empty spaces between them (Fig. 11) thatwas not seen by thin-section TEM. x,z scans through similarregions showed a diversity of infrastructure within this meshwork(Fig. 12A and B) and provided a wealth of dimensional data.Individual fibers, as small as 8 to 14 nm in diameter, couldbe distinguished (Fig. 12B). If allowance is made for tip convolutioneffects, these individual fibers may have widths in the orderof a few peptidoglycan strands. The spatial resolution in ourexperiments was, however, insufficient to clearly identify anddetermine the orientation of individual structural componentsof the cell surface. The diameter and depth of the depressionswithin the network were variable but were usually on the orderof 2.5 to 4.0 nm (Fig. 12C and D). It is likely that the measureddepth was limited by tip geometry. The structures shown in Fig.11 and 12 were observed over the contoured regions of the cellsurface (Fig. 13A, regions a) where negligible tip adhesionwas observed. Reproducible imaging of the sponge-like structurecould be obtained for different scan speeds, scan directions,and imaging forces between 0.5 and 3 nN. We therefore believethat the images reflect an intrinsic structure of the cell surfaceand not a tip induced artifact.

View larger version (203K):
[in this window]
[in a new window]
FIG. 11. High-resolution AFM image of an older region of the cell wall. The image shows a network of fibers with large empty spaces. Bar = 50 nm.

View larger version (58K):
[in this window]
[in a new window]
FIG. 12. (A) High-resolution contact mode AFM image of an old wall section. (B) Cross section measurement along the line shown in panel A. Individual fibers as small as 8 to 14 nm can be identified. If allowance for tip convolution is made then such fibers could have widths on the order of a few peptidoglycan strands. (C) High-resolution image of another section of an old cell wall. (D) Cross-section along the line shown in panel C. The depth of some of the depressions in these images were 40 nm and depth measurement was limited by AFM tip geometry.

View larger version (71K):
[in this window]
[in a new window]
FIG. 13. (A) Deflection image of an older section of the bacterial wall. The image shows contoured regions (a) as well as much smoother areas (b); Bar = 40 nm. (B) Force spectra over the contoured regions (a) showed negligible tip adhesion when cantilevers were retracted from the surface. Cantilever retraction curves over the smooth regions, however, showed tip adhesion effects (see double-headed vertical arrow [curve b]), and a force spectrum characteristic for polymer stretching was generally observed. Rupture length, indicated by the horizontal double-headed arrow, was 200 nm.

On a coarser scale, these older cell wall surfaces often showedconsiderable variation in the way the AFM tip interacted withit. Deflection images revealed that the surface was actuallymade up of smooth and contoured zones (Fig. 13A). Force curvescollected over smooth zones (Fig. 13A, regions b) showed thecharacteristic force spectrum expected for polymer adhesion(36). After adhesion and during tip retraction, the polymersin these smooth zones stretched until rupture. The maximum stretchingforce (Fig. 13B, curve b) was $~$ 300 ± 50 pN and the rupturelength is about 200 nm (Fig. 13B, curve b). On the other hand,the more-structured contoured zones only showed force spectracharacteristic of nonadhering tips (Fig. 13B, curve a); here,the tip approach and retraction curves were identical. Clearly,there must be considerable physicochemical differences betweenthese two types of zones that are due to their macromolecularcomposition. Each zone could reflect the degree of surface autolysisor secondary polymer concentration in these regions of the wall.A detailed study of the variation of adhesion force over thebacterial surface is in progress and will include adhesion mapping.

General comments on cell division. Bacteria, especially gram-positive varieties, must be carefulduring growth (i.e., the expansion of existing cell wall) anddivision (i.e., septation and cell splitting) since they areunder considerable turgor pressure. Uncontrolled wall turnoveror septation could deregulate autolysins resulting in cell lysis.For S. aureus, septation and cell splitting (i.e., the actualseparation of one daughter cell from the other) are importantevents since much of the new wall material is incorporated andremolded at these stages for each cell. According to Giesbrechtet al. (22, 23), cell splitting starts with the creation ofa row of holes in the peripheral cell wall directly above theseptum (or cross-wall). These holes are thought to be structuresin the wall (i.e., murosomes) possessing high autolytic activity,which digest away adjacent peptidoglycan and lead to the splitting.More recent data suggest that a bifunctional autolysin withboth amidase and glucosaminidase activity (Atl) forms an initialring structure on the surface at the septal site (6, 32, 37,43). It is possible that the holes shown in Fig. 8A are theresult of this early autolytic activity eventually producingGiesbrecht's murosomes. Our AFM images show the holes to be50 to 60 nm in diameter and to have a center-to-center spacingof $~$ 100 nm. For S. aureus with a diameter of $~$ 1 µm, thisimplies about 30 perforation holes initially around the bacterialcircumference. As our AFM imaging proceeded, the holes becamelarger and merged into one other (Fig. 8B and C), thus creatinglonger cuts in the peripheral wall as splitting proceeded. Thisaction by murosomes would continue from the septum peripheryinwards until the daughter cells had completely split. Thin-sectionTEM showed a heavily stained midline in the septum below themurosomal action (Fig. 3C and D and 6), which suggests thatpeptidoglycan hydrolysis in septa is already occurring at theseregions in advance of murosome activity, possibly to make roomfor the advancing murosomes and their splitting activity. Asmurosomes perforate the septum, tension will now develop inthe two halves of the cross-wall, and cell separation, in principle,will continue through purely mechanical means provided the twohalves of the cross-wall are not too strongly bonded along themidline of the splitting system.

Once the daughter cells have separated, their new walls (thatwere previously part of the cross-wall) will become accessibleto AFM imaging. Here we see the concentric rings that have beenseen by other researchers via scanning EM and TEM (4, 5, 8).The question arises as to what this ring pattern actually represents.One possibility is that the ring structure represents structuralfeatures of the splitting system created by a synthetic processthat produces alternate ridges and grooves. The above-mentionedTEM studies were carried out with S. aureus, Staphylococcusepidermis, and Bacillus anthracis and show the rings on boththe inside and outside faces of the cell wall. This stronglysuggests to us that the ring pattern is an integral part ofthe fabric of the new cell wall and that it most likely doesnot only represent structural features of the splitting system.Since new peptidoglycan is inserted at the inner face of thewall via penicillin-binding proteins (33) and since the splittingsystem must involve autolysins and the breakdown of peptidoglycanas the daughter cells separate, it is logical to assume thatthe rings are structural features or remnants of the puttingin and taking out of the peptidoglycan during the division ofthese gram-positive bacteria. For example, if the peptidoglycanstrands are oriented perpendicular to the cross wall (see nextparagraph), then some circular regions may have the peptidoglycanends protrude slightly above the plane of the cross-wall whileother circular regions might have the ends slightly below thisplane, thus generating a landscape of circular grooves and valleys.Peptidoglycan strands are composed of disaccharide units $~$ 1 nmin length, and a single tessera of the peptidoglycan wall fabrichas a characteristic length of $~$ 4.5 nm (29). These dimensionsare comparable to the dimensions of some of the topographicfeatures of the ring structure discussed above and it is possiblethat the topographic features of the rings reflect structuralfeatures of the peptidoglycan strands.

The concentricity of the rings implies that the peptidoglycanstrands that help make up the wall are, themselves, orientedin a similar fashion and are added or removed in an orderlymanner. There could be three distinct ways of doing this: (i)the glycan strands lie in the plane of the cross-wall and areoriented radially, (ii) the strands are oriented tangentiallyto the inner surface, or (iii) the strands are oriented perpendicularto the plane of the cross-wall. In cases i and ii, the peptidoglycanstrands would end up lying in the plane of the cross wall andhence in the plane of the new cell wall after cell separation.This would imply that after several cycles of division in S.aureus, where new cross-walls are approximately orthogonal tothe last division plane, the cell wall would break up into amosaic of peptidoglycan orientations that would weaken the overallbonding in the wall (i.e., similar to crystal defects in othernetworks, such as cellulose), thereby affecting the wall's strength.The cell wall has to support a large turgor pressure (28), andit is not clear that this in-plane arrangement of peptidoglycancould provide the necessary mechanical strength to contain thispressure (34). It therefore seems that the third possibility,in which strands are oriented perpendicular to the cell wallto form a scaffolding, may be the preferred option. The possibilityof a scaffolding-like peptidoglycan arrangement in gram-positivewalls has recently been suggested by Dmitriev et al. (16, 17),but clearly more definitive data beyond what we have found herewill be necessary for this model's confirmation.

In a three-dimensional peptidoglycan network, as exists in theS. aureus wall, all peptide stems (i.e., the peptide portionof the glycan strand) can, in principle, participate in cross-linkingand a complex multilayered wall structure is possible. UnlikeEnterococcus hirae, where all new wall material emanates fromthe cross wall (24), some turnover of preexisting wall in S.aureus is possible (13, 23). Accordingly, old cell wall materialshould be shed from the outer wall surface, making it less cross-linkedand less tightly bound than the underlying regions. It was,therefore, interesting that preexisting wall had a loosely arrayednetwork on its surface, consisting of fibrils and holes (Fig.11 and 12A to D). Indeed, this loose meshwork resembled a gel-likematrix that would be highly hydrated; our impression is thatthe fibrils consisted of linear aggregates of two or more peptidoglycanstrands of low bonding order. At this time it is impossibleto determine if the strands would be planar (Koch's tesseramodel [27]) or perpendicular (scaffolding model of Dmitrievet al. [17]) to the flat axis of the surface since we couldnot image individual strands by AFM because of tip convolutioneffects. Yet, our gel-like topography would be consistent withactive hydrolysis of the surface by autolysins and wall turnover.

Differential AFM provided better contrast to the old wall surfaceand revealed that it did not entirely consist of the gel-liketopography. Some regions were relatively smooth (Fig. 13A).Force curves of both zones (Fig. 13B) revealed that the smoothzones possessed more adhesiveness than the gel-like zones, suggestiveof macromolecular differences or rearrangements between thezones. It is possible that the more loosely arranged zones (Fig.13A, region a) are soft zones with little structural integritybecause of wall turnover. They would therefore have no stretchingcapacity once stuck to the AFM tip because of reduced polymerbonding and this would be consistent with our explanations above.

In summary, the data produced by AFM on S. aureus correlatewell with the thin-section TEM data and considerably expandit. AFM was able to follow the division process with precisionand confirm the initial splitting of the septum, presumablyby murosomes. After daughter cells had separated, concentricrings were found on the surface of the new wall. The old wallcontained two surface zones, a smooth zone that had a high degreeof adhesive properties and a gel-like zone that was less adhesive.These results on division, new walls, and old walls are interpretedat the macromolecular level so as to implicate peptidoglycanassembly and turnover and are consistent with either the planaror scaffolding models of peptidoglycan arrangement.

ACKNOWLEDGMENTS

We are indebted to V. Matias for the processing of the S. aureusculture for thin-section analysis, D. Moyles for thin sectioning,R. Harris for assisting with image formatting, and A. Saxenafor the supplying the culture (all four are members of T.J.B.'slaboratory).

This research was supported by an NSERC-CIHR collaborative grantto M.J. and T.J.B. The EM was performed in the NSERC GuelphRegional STEM Facility, which is partially supported by an NSERCMajor Facilities Access grant to T.J.B.

FOOTNOTES

* Corresponding author. Mailing address: Department of Physics, Dalhousie University, Halifax, Nova Scotia, Canada B3H 3J5. Phone: (902) 494-2316. Fax: (902) 494-5191. E-mail: jericho{at}fizz.phys.dal.ca.

REFERENCES

Top